IBM Research Africa, Set to Solve Real-World Problems

By Uyi Stewart
Uyi Stewart, Chief Scientist, IBM Research – Africa
Uyi Stewart, Chief Scientist, IBM Research – Africa

In an interview with Wired magazine, the English musician, Brian Eno, complained that there is not enough Africa in computers.

How does one Africanize…or otherwise liberate a computer?” he wanted to know.

Maybe Brian would like to visit us at our new research laboratory in Nairobi, because this is more or less what we are doing. Although our focus is not to build computers, per se, we are building technology solutions for Africa— with uniquely African flavour. Africanized solutions, if you like.

IBM Research—Africa, officially opens its doors next week. It’s our 12th global research laboratory, and the first in Africa. It feels like a pivotal moment. It certainly is for me.

As the lab’s chief scientist, it is my job to develop solutions for what we call Africa’s grand challenges. The grand challenges that we will focus on, across Africa and starting here in Nairobi, are education, water, energy, healthcare, financial inclusion, public safety, human mobility and agriculture.

Although we are just opening the doors on the facility itself, as an organization, over the last year we have been busy addressing a number of these challenges. The first to be launched is a proof-of-concept which we hope will make commuting in Nairobi a bit less painful than it currently is.

Much of Nairobi’s road network is more than half a century old and was developed for a city of just 350,000 inhabitants. Like most African cities, Nairobi has been hit by rapid urbanisation. Let’s just look at cars: today, there are two million cars on Nairobi’s roads and 400 more cars are added to the city’s already congested streets every day. It is estimated that traffic congestion costs the city Sh50 million (USD $600,000) in lost productivity every day.

Last year, our researchers identified human mobility and the related issues of transportation and public safety as one of Nairobi’s most pressing challenges for future growth. It was clear that, as a matter of priority, we needed to develop a system to ease traffic congestion.

With our partner, internet service provider Access Kenya, we have developed a mobile phone service that provides drivers with live updates on traffic jams and suggests alternative, less congested, routes.

How did we do it? We took some existing bits and pieces, collaborated, used our ingenuity, and built an Africanized solution. We started with a transportation model that was developed at our lab in Tokyo. What we lacked in Nairobi, though, were data feeds.

When we noticed that Access Kenya had 36 webcams, streaming live images of traffic in Nairobi’s central business district (CBD), we approached them and asked if we could use their data. They chuckled and said, “What are you talking about? No one has ever come to us.” But they could see our vision and agreed to work with us.

The webcam images are low resolution, so the first thing we did was build image-enhancement software to make the vehicles stand out more. Once we had done that, we were able to work out how fast vehicles were moving and also the levels of congestion. But we were initially stymied by the fact that we only had data feeds from 36 cameras, all located in the CBD.

In order for this data to be meaningful, we needed to be able to generalise the traffic situation across all of Nairobi. So the second thing we did is—and this is the really smart bit—we built an “inference model” that take a sub-set of data from the CBD and generalise it across all of Nairobi.  This is computer science at its best. The stuff I get out of bed for.

Based on these home-grown innovations we are now able to provide motorists with information about traffic congestion in Nairobi through SMS, and for those who have smartphones, a map. The idea is that commuters will check the traffic situation on their phone before heading to their destination.

The service has just been launched and is available as a trial for all Safaricom and Airtel customers (Safaricom customers must enter the short-code *384*3# to access the service, and Airtel customers,*381#).

We deliberately didn’t over-engineer the solution. We just wanted to get it out there because the more people use it, the more the system learns, and the better it gets. Our researchers are currently working to extend the capabilities of the solution to include data on pubic safety, weather conditions and road works to create a Nairobi-specfic view of human mobility

This is just one example of how we are taking on the grand challenges and how we can build smarter cities in Africa without breaking the bank. Our approach is frugal, pragmatic, collaborative and ambitious. But above all, it is smart. And my goodness! If there’s one thing Africa has in abundance, it is smart people.

Virtual Recruiting Event

Scientists from IBM Research – Africa will be hosting a live virtual job fair in the IBM SmartCloud and Google+ on 05 December to talk about several open positions. For more details, visit IBM JobFair.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.